Write an Input Adapter

from beekeeper import EntityInputAdapter


class CsvWorkerAdapter(EntityInputAdapter[MyWorker]):
    def __init__(self, file: Path) -> None:
        self._file = file

    def get_entities(self) -> Iterable[MyWorker]:
        with self._file.open() as f:
            for row in csv.DictReader(f):
                yield MyWorker(name=row["name"], role=row["role"], unavailabilities=[])

For allocations, subclass AllocationInputAdapter[YourRequest] and implement get_allocations(self) -> Iterable[YourRequest].

Most domains will have one of each. Compose them with CompositeInputAdapter:

adapter = CompositeInputAdapter(
    entity_adapter=CsvWorkerAdapter(Path("staff.csv")),
    allocation_adapter=ApiAllocationAdapter(api_url="..."),
)

If both kinds of data come from the same source, subclass InputAdapter[TEntity, TAllocReq] directly and implement both methods on one class.

For JSON-from-file, the framework ships JsonEntityInputAdapter and JsonAllocationInputAdapter — strict-only, pydantic-backed, no leniency knob. Use them or write your own.